Chapter 2. War, the First Phase: September 1939–April 1940

Both Britain and Germany made important preliminary moves before fighting broke out on 1 September with the invasion of Poland. The RN’s Home Fleet, which had been exercising from Invergordon, now moved to its old-time base at Scapa Flow, still largely unprotected from attack. A well-timed review of the Reserve Fleet had been held in Weymouth Bay during August, ensuring that the old destroyers of the V&W classes would have crews and that their machinery was tested and ready.

British operations were directed by the C-in-C Plymouth, Adm Sir M Dunbar-Nasmith, VC, an outstanding submarine commander of World War I. Like so many senior service officers of democracies at the opening of a war, he had to tackle a difficult task with totally inadequate resources. However, he founded the training organisations at Liverpool and Tobermory (see chapter 3) and planned the move of Western Approaches Command to Derby House, Liverpool. His successors owed much to his preparatory work.

PC74 was built in World War I as a P boat with merchant-ship appearance. She was used as the decoy ship Chatsgrove in early World War II, with no success, before reverting to auxiliary duties. (IWM A 25975)

Germany had fifty-seven U-boats – almost exactly the intelligence estimate – on the outbreak of war, a very small proportion of the 300 Atlantic boats that Dönitz saw as needed to win a war on merchant shipping. Of these boats, thirty-nine were operational, the remainder being used for training or working up after completion. Another six were due to complete by the end of 1939. Between 15 and 19 August eighteen U-boats sailed for their war stations – five small Type IIC and thirteen Type VIIA or B. This was a maximum-effort deployment before the British were able to introduce counter-measures such as convoy. The ‘pocket battleship’ Graf Spee sailed from Wilhelmshaven on 21 August, followed by the Deutschland three days later, accompanied by their supply ships Altmark and Westerwald. The Admiralty saw these surface raiders as a greater threat than the submarine.

In these early days Hitler still hoped for a short war against Poland, followed by a negotiated peace with Britain and France. In consequence, Dönitz was ordered to operate strictly within the Prize Rules, forbidding attacks on merchant ships without warning. Unfortunately for any such hopes, Lemp in U-30 sighted a liner in poor visibility that, from its position and movements, he thought wrongly to be an armed merchant cruiser. He fired two torpedoes, both of which hit and exploded, an unusual success rate. He had sunk the Donaldson liner Athenia of 13,851grt, carrying some 1,100 passengers, of which over 300 were Americans. Despite prompt rescue attempts 112 were drowned, including 28 Americans. The Nazi command made frantic efforts to conceal the involvement of a German submarine, while the British drew the parallel with Lusitania. Propaganda apart, the government and the RN saw this as a clear indication that Germany intended to wage unrestricted submarine war from the start. It is probable that this misinterpretation made little difference, as submarine warfare ‘to the rules’ is so difficult that an accident was almost bound to happen. Initially, the German command re-emphasised the need to observe the Prize Rules, but on 24 September Dönitz ordered that ships using their radio should be sunk without further warning.

As discussed in the previous chapter, the Admiralty had well worked-out plans for convoys, though it was not intended to implement them until it was clear that Germany had adopted unrestricted warfare. The sinking of Athenia was taken as such confirmation and convoys began to operate. One convoy had jumped the gun: eight tankers, bound for the Gulf, had been held at Gibraltar until Italy’s intentions became clear. They left on the long trip round Africa on 2 September under cruiser escort as far as Freetown.

Convoy OA1 of seventeen ships bound for North America left the Downs on 7 September, while OB1 of ten ships left Liverpool the same day. Coming the other way, HX1 of fifteen ships left Halifax, Nova Scotia¹ on 16 September, followed by a fast convoy, HXF1, on the 29th. The speed of HX convoys was set at nine knots, though eight knots was more typical of that achieved. HXF convoys were for ships that were faster but could not achieve the fifteen knots set for independents. The speed for SC convoys, slow passages from Cape Breton, was set at seven and a half to eight knots. These early convoys were small in numbers of merchant ships but within a month or so built up to the thirty to forty ships then thought optimum. The times of passage for the main routes were:

HX 15.2 days from New York

SC 15.4 days from Sydney, Cape Breton

SL 19 days from Freetown

Under Dunbar-Nasmith, many of the measures that would lead to victory were initiated. The escort force for these convoys was often only a single destroyer and a sloop, with little training or experience in ASW, which remained with the convoy only as far as 15ºW. Fast ships sailed independently, as it was thought that a submarine with a maximum surface speed of about seventeen knots, reduced in bad weather, would not be able to attack. The lower limit of speed to sail independently was set at fifteen knots, reduced to thirteen in November 1940. During September, forty-eight merchant ships were sunk by U-boats, of which forty-three were sailing independently, and only one from convoy (the others were mined). Lowering the qualifying speed made matters dramatically worse. On the Halifax route, the chance of an escorted ship’s being sunk on a round voyage was 5.8 per cent, whilst that for an independent (thirteen to fifteen knots) was 13.8 per cent.

Various attempts were made to remedy the shortage of escort vessels. The idea of ‘Q’ ships, which had some success in World War I, was resurrected. By March 1940, eight decoy ships were fitted out with up to nine 4in guns, torpedo tubes and depth charges.² They were to simulate a straggler and catch a U-boat unawares, and even to take on a surface raider with some prospect of causing damage. None of them had any success and two were lost. An enquiry was held in December 1940, as a result of which the scheme was abandoned.

The Admiralty was still worried by the potential threat from surface raiders and by February 1940 forty-six armed merchant cruisers had been commissioned. Most convoys in more distant waters had an escort of a single such ship, though since most had obsolete guns, little or no fire control and crews made up of reservists and National Service youths, their prospects in a duel with a raider were not good.

The Admiralty, egged on by the First Lord, Churchill, had fallen for the old fallacy that escorting convoys was ‘defensive’ and that we should go on the ‘offensive’ with hunting groups, ‘like cavalry’. A pre-war study by the Tactical Division of the Staff had proposed nineteen groups, each of five ships with air support (some from fleet carriers). With only about 100 escort vessels available, this would have left few for convoy work. Some hunting groups were formed, a very few including a fleet carrier. A comparison of the one-mile effective range of asdic with the 3,000-mile expanse of the Atlantic Ocean should have shown the folly of this approach. Later, with submarine concentrations localised by Enigma and plenty of escorts, hunting made sense.

Gardenia was a very early Flower and unusual in completing with two masts.

On 14 September U-39 sighted Ark Royal off Orkney and fired two torpedoes from an ideal position at a range of 800 yards. Both exploded prematurely some eighty yards short. This revealed the position of U-39, and the destroyer Foxhound leapt to the attack, blowing the submarine to the surface and capturing her crew in the first U-boat sinking of the war. The next day, two Skuas from Ark Royal attacked U-30, causing superficial damage, but the explosion of the bombs before reaching the sea caused both aircraft to crash. Two of the aircrew were rescued by the submarine.

On 17 September U-31 sighted the carrier Courageous, which was hunting submarines with four destroyers. Two of these were detached to investigate a visual sighting by one of her aircraft some seventy miles from the carrier leaving only two destroyers as escort and planeguard. A change of course put U-31 in an ideal attacking position at a range of about 3,000 yards. She fired three torpedoes, of which two hit, sinking the carrier with the loss of 518 men – and one of the RN’s five carriers.³ This put an end to the practice of using fleet carriers to hunt submarines but it was another year or two before the Admiralty stopped taking escorts from the screen to prosecute a contact some hours’ steaming away.

A month later, on 14 October, Prien took his U-47 through an unblocked channel into Scapa Flow and sank the battleship Royal Oak. Prien was a brilliant and dedicated U-boat officer and a fanatical Nazi, who would become one of the most successful U-boat commanders of the war. He had visited Orkney as a ‘tourist’ before the war and studied the entrances to Scapa Flow.

Coastal Command of the RAF was unable to offer much in the way of air support. Of a force of seventeen squadrons, only one squadron of Hudsons and two of Sunderland flying boats had sufficient endurance to operate over the Atlantic convoy routes. The majority of the Command’s aircraft were Ansons, with insufficient endurance even to reach the coast of Norway, and a load of four ineffective 100lb bombs. To be fair to the Air Ministry, they had been told by the Admiralty before the war that the submarine threat had been neutralised and that the main threat to shipping came from surface raiders. An anti-submarine bomb had been developed, starting in 1925, which entered service in 1931. An ‘improved’ Mk III was adopted, without trials, in 1934. There were three sizes – 100lb, 250lb and 500lb – all suffering from a fuse which was virtually useless and an unpredictable underwater trajectory. In September the British submarine Snapper was hit by a 100lb bomb dropped by an Anson. The bomb hit at the base of the conning tower, causing no structural damage but breaking four light bulbs. The 250lb bomb was claimed to be lethal at six feet from the target, and the 500lb at eight feet.

Coastal Command’s greatest asset in the early days was its commander-in-chief, Sir Frederick Bowhill. He had entered the Navy in 1904, transferring to the RNAS and following it into the RAF in 1918. He understood the ways of ships and the sea but with an airman’s eye.⁴

Problems

During a war, staffs tend to underestimate the problems of their enemy and this fault is even more common amongst the works of later analysts, despite the benefit of hindsight. During the opening phase of this war, both sides had plenty of problems. Common to both was a shortage of vessels with which to implement their plans. Dönitz had more direct problems. The engines of the Type VII seem to have given trouble and their engine mounts needed stiffening urgently. These problems were not serious but exposed weakness in the refit organisation.

The torpedoes fired by the U-boats had much more serious problems.⁵ In the 1930s Germany had developed two types of torpedo: the G7a and the G7e. The former was a good, conventional torpedo running on compressed air and Decalin (decahydronapthalene). It had a 300kg warhead and three speed settings up to forty-four knots, at which it had a range of 6,000 metres. The forty-four-knot setting was temporarily deleted early in the war but reinstated after the engine was modified.

The G7e had an electric drive with a battery weighing 665kg, which gave a speed of thirty knots for a range of 5,000 metres; but there was no bubble track. It was developed in great secrecy at Karlskrona (Sweden) from 1929 and its existence was unknown to the RN until remains were found close to the wreck of the Royal Oak in Scapa Flow. Full details came with the capture of U-570 in August 1941. Other than the lack of a bubble track, the great advantage of the electric torpedo was that it was easier and cheaper to manufacture. In 1939 a G7a required 3,730 man-hours, reduced to 1,707 by 1943, whereas the G7e needed only 1,255 man-hours. Production of all types of submarine-launched torpedoes (mainly G7e) was at the rate of seventy per month in 1939, rising to 1,000 a month by early 1941, peaking at 1,700 by 1943 and falling back to 1,400 in 1944.

The first problem to become apparent was in depth-keeping, with frequent examples of torpedoes running up to 3.7 metres too deep. This problem had been recognised before the war but was not taken too seriously, as it was expected that the magnetic firing pistol would normally be used, and thus a small depth error would not matter. The problem was found to be a build-up of pressure in the balance chamber controlling depth-keeping through the hydroplanes. This was due to leakage from the air inside the submarine, which inevitably increased in pressure during submergence due to high-pressure air leaks, inboard venting, etc. Once this was diagnosed, better seals were provided and the problem was solved.

The MZ magnetic pistol had its own problems. It was oversensitive to changes in the perceived magnetic field and could be triggered by rapid manoeuvres of the torpedo and by local magnetic anomalies, which were common off Norway. Degaussing, as used by the British as protection against magnetic mines, also much reduced the distance at which the pistol would function. There was no easy solution and the magnetic pistol was deactivated until the end of 1942. This, in turn, exposed the failings of the contact pistol, which did not function when hitting at an oblique angle. RN magnetic pistols had similar problems but the fallback contact pistol was very reliable – so much so that the Germans considered copying the one they captured in Seal.

Mine Warfare

The RN soon closed the English Channel by laying 10,000 mines in both shallow and deep fields supported by patrols. The mines were of the World War I ‘H II’ type, but still effective.⁶ After losing three Type II boats to these mines, Dönitz forbade the use of this route, so that all U-boats had to go round the north of Scotland. The Type IIs had insufficient endurance and operated in the North Sea, where they were exposed to the East Coast Barrage. An attempt to lay a new Northern Barrage was unsuccessful and uncompleted when the Germans invaded Norway. There was an attempt to mine the route to Greenland and Iceland, which was also unsuccessful, as was a field in St George’s Channel.

On the other hand, mines laid by U-boats had some notable successes. The cruiser Belfast broke her back and suffered severe shock damage on a mine laid by U-21 on 21 November,⁷ while Nelson was badly damaged, with many casualties, by a mine from U-31 on 4 December 1939. (Nelson suffered a large number of casualties from broken porcelain toilet pans.) Pre-war trials had shown the problems of shock damage but there was no time for remedial action before war broke out. These trials were conducted on a short pontoon and failed to show the effects of whipping. Up to the end of February, thirty-seven merchant ships totalling 129,000 tons had been sunk by submarine-laid mines.

Phase 1 of the Atlantic battle. The pre-war view that attacks would take place mainly in or close to home waters was justified, the southwestern approaches being the main danger zone.

The Balance Sheet So Far

Both sides had grounds for reasonable satisfaction on the way things had gone, but both had continuing problems and entertained hopes for better things to come. For both Britain and Germany the outstanding problem lay in shortage of numbers. Dönitz had asked for twenty-nine new boats per month but was forced to accept a target of twenty-five. It made little difference, as during the first half of 1940 only two boats were delivered each month, rising to six in the second half of the year. U-boat building was given priority from mid-1940. This was not enough to make up for losses, and the number of boats available for operations fell from thirty-eight in December 1939 to twenty-five in the following November. The number on station was remarkably constant at one-third of the total, say six to eight.

On the British side, matters improved slightly as hunting groups were abandoned, making more vessels available for convoy escort. The only new escorts coming forward in numbers were the Hunt class destroyers, whose endurance was inadequate for Atlantic duties, and the Flower class corvettes, considered in the next section.

Up to May 1940, twenty-four U-boats had been sunk (see table 2.1). All the sinkings by surface ships resulted from depth charge attack. One was by an old sloop, one by a minesweeper and the rest by destroyers.

Surface vessels	13½*
Aircraft	2½
Submarine	3
Mines	4
Accident and unknown**	1
Total	24

* ½ indicates shared

** In this and later tables it is reasonable to assume that some losses with unknown cause were due to mines

Sinkings of merchant ships by U-boats fluctuated considerably, mainly because of the small numbers involved. In the months up to March 1940, when the U-boats redeployed for the Norwegian campaign, the average tonnage sunk per month was some 160,000–180,000grt, of which 60 per cent were ships sailing independently. This represented 57,000 tons per U-boat sunk, nearly as low as in 1918 when the first U-boat war was lost. In the first few months of the war, losses of merchant ships had been made good by some new building and a ‘one off’ bonus of captured German shipping. Dönitz was not unduly worried, as things could only improve as more U-boats entered service. This was the war the Admiralty had planned for and, by and large, they got it right.

Asdic in Use

The introduction of asdic⁸ led to a major change in anti-submarine warfare (ASW): for the first time a fully submerged submarine could be detected and located with sufficient accuracy for weapons to be launched with a reasonable promise of success. It is no reflection on the pioneers of asdic to point out that sets in use in 1939 had a number of limitations, some serious. Some of these problems were eliminated during the war by improved hardware, while the effect of others was reduced by improved operational methods. Most of the problems were known to specialists before the war and action was in hand. However, to many officers, asdic was a wonder weapon and knowledge of its limitations came as a great shock.

The transducer generated a short and narrow pulse of very high frequency sound, which would be reflected back if it hit a solid object. To aid interpretation, the pulse was represented by audible-frequency sound, the outgoing pulse emanating from the loudspeaker as a ‘ping’. The beam was quite narrow – 5º – and the operator had to wait for the reflection before moving to the next 5º sector. It could take a long while to search a wide arc and the submarine could move a considerable way between pulses. More numerous escorts reduced the arc that each had to search, while later sets could search more quickly.

This composite drawing by John Roberts shows (a) The conical beam of the main set (120 series), effective out to 2,500 yards in good conditions but with no ability to measure depth. It was not able to detect submarines at great depth.

(b) The Q attachment, introduced to overcome these problems. It had a very narrow (3-degree) beam in the horizontal plane and only needed a very small strip below the main transducer.

(c) The 147 was a more advanced depth measuring set and was used to set the required depth on Squid projectiles immediately before firing.

See also Chapter 7 Page 114.

Asdic dome for Type 121 and typical of all early retractable domes. (John Roberts)

If an echo was received, the range and bearing could be read off, the bearing recorder giving something of a moving image. These would be transferred to the plot, together with any data from other escorts, and the course and speed of the target determined. With early wartime asdics, the average detection range was about 1,300 yards, with 2,500 yards on occasional, good days. If the contact was stationary, it was probably not a submarine unless bottomed in shallow water. A moving echo could be received from a whale or a shoal of fish or even tide rips, though skilled operators could usually tell the character of the echo. The plot would distinguish between the erratic movement of fish and the purposeful movements of a submarine.

Water moving over the face of the transducer could generate noise, so from early days the transducer was carried within a water-filled, streamlined dome so that the water in contact was not moving. The dome was of light construction so that it did not obstruct or distort the sound beam, but this meant that it could not be exposed to the water pressure at high speeds. In consequence, the dome was made retractable, which gave serious problems. Ideally, it should be on the centre line, which meant cutting away the keel over a considerable length, but available structural design methods were unable to cope with such a discontinuity in the area of peak slamming pressures. In slow-speed ships, retractable domes were not needed and the dome could be removed in dock. Such domes were often referred to as ‘trawler sets’. Asdic was so secret that its importance was not explained to the ship designers, a common problem even today.

The sound beam spread in a cone shape from the transducer with the axis of the cone at an angle below the horizontal. This gave the only indication of target depth. The deeper the submarine, the greater the range at which it would enter the leading edge of the beam. As the escort got close to the submarine, the target would come out from the rear of the beam giving a further indication of depth but adding to the dead time before weapons – depth charges – could reach the target. The dead time would be from the time contact was lost till the stern depth charge rails were over the submarine, together with the time for the depth charges to sink to the estimated depth.

The Mk VII depth charge, basically the Type D of World War I, sank at an initial rate of 7ft/sec, accelerating to 9.9ft/sec at 250 feet. It weighed 429lb and carried a charge of 290lb of amatol (later minol; TNT filling was used in hot climates), which it was thought would be lethal if exploding within thirty feet of the submarine. In fact the lethal distance was more like twenty feet, though an explosion forty feet away might force the U-boat to surface. Since depth estimation was only slightly better than guesswork, the chance of a successful attack was remote; in the first six months of the war 4,000 attacks were made, sinking thirty-three U-boats – less than 1 per cent. By 1943–4, better training and improved equipment brought the chance of a successful depth charge attack to about 5 per cent.

The Mk VII depth charge was little changed from the Type D of World War I. Initially its charge was 290lb of amatol, but this was later changed to minol. Complete, it weighed 429lb. A faster-sinking version, Mk VII Heavy with an additional 150lb weight, was later introduced – as shown as an insert in the diagram. (John Lambert)

The original depth charge thrower, the Mk II (top right and far right). The carrier and stalk were thrown with the charge, which was wasteful in weight and storage space. It was later replaced by the Mk IV (centre right and far right) in which the carrier and stalk were retained. (John Lambert)

Depth charge stowage rack. (John Lambert)

The dead time is a key aspect of antisubmarine warfare in World War II. The sinking time to 250 feet was about thirty-five seconds, to which must be added some twenty to twenty-five seconds for loss of contact (typically 200 yards), a total of nearly a minute. At four knots (5ft/sec) a submarine could move 300 feet in any direction, far in excess of the lethal distance of some thirty feet. The deeper the U-boat, the longer the time required for the depth charge to sink and the less likely a kill. The seriousness of this problem had been partly concealed in prewar exercises in the Portland area, where the sea is fairly shallow.

The sea is not uniform in density. Differences in salinity and temperature cause differences in density, which lead to changes in sound transmission that can bend a sound beam as a lens or prism bends a light beam.

This diagram by John Lambert shows the complex electric and hydraulic systems needed to fire a ten-pattern charge under asdic control.

The basic depth charge pattern was the diamond, with a charge at each corner and one in the middle, all set to the same depth. Initially the charges were forty yards apart, but after studies at Fairlie this was increased to sixty yards in 1943 when amatol was replaced by minol. In late 1940 the Mk VII Heavy was introduced, with a 150lb weight attached, which sank at 16ft/sec and had a minol charge, lethal at twenty-six feet and disabling at twice that distance. Using two rails and four throwers it was possible to drop a total of ten charges, mixed heavy and light, which would explode as two diamond patterns, fifty feet apart in depth – the ‘ten pattern’. The full pattern for a mean depth of seventy-five feet began with a heavy charge from the rail followed three seconds later by two heavies from the forward throwers and a light from the rails. After a further eight seconds, the after throwers discharged two lights with a heavy from the rails, and then the remaining charges from the rails. Later, a fourteen pattern was tried, but with only a total of 100 charges or less carried in most ships the ten pattern was seen as the best compromise. Initially the ex-Brazilian Hs had eight throwers and three rails, capable of launching seventeen charges.

The one-ton Mk 10 was introduced in December 1942. With a charge of 2,000lb, it was said to be as effective as a ten pattern of Mk VIIs and better at depth. It was fired from a torpedo tube in destroyers and a special launching rail in the Captains. Early versions sank at 6ft/sec to give time for the dropping ship to get clear, later increased to 21ft/sec with a depth setting of 640 feet. It was determined that the ship should be safe at eleven knots with a depth setting of 220 feet. The weapon seems to have been used only rarely. The original thrower Mk II was replaced by the Mk IV from September 1941, which retained the arbor on which the charge rested.

This is the meaning of ‘lethal damage’: the effect of an underwater explosion on the submarine Stygian in a post-war trial.

Flower Class Corvettes

By 1938 it was clear that there were too few escort vessels, both for ocean work and, more urgently, for the east coast. The points made in the review of 1932 remained valid but the threat was clear and growing, as were the problems. Rearmament had started but this made matters worse for the building of numerous escorts, as the traditional builders of ships and engines were all working to capacity and beyond.

Two memoranda, summarised here, by Sir Stanley Goodall in early 1939 shed light on options for cheap A/S vessels as war approached. Consideration had already been given to the use of commercial whale catchers. These needed much more alteration than trawlers because of their poor subdivision and there must be some doubt regarding the quoted cost of £50,000, which may have referred to 1936 when prices were much lower. Alterations to convert existing ships would be extensive and costly.

Experience in the First World War showed that the requirements for A/S vessels increased in both size and speed from 42-ton MLs to twenty-two-knot, 573-ton P boats. USN experience was similar in starting with 110-feet submarine chasers before the 590-ton Eagles with a speed of twenty-two knots. Speed is expensive but it is an unwise economy to build slower ships than the submarines capable of eighteen knots on the surface that are being built by hostile powers.

Two factors of importance to consider in connection with the type of A/S vessels contemplated for construction during wartime are the source of production and the time for building. These two factors are inter-connected. Ships such as the Admiralty design based on trawler practice could be built rapidly because the machinery is simple and the yards concerned are merchant ship builders specialising in this type of work; whereas ships like patrol vessels [Kingfishers] with more complex machinery should preferably be built in yards that would be busy with other warship work.⁹

MTBs fitted for A/S work [MASB] seem most suited for local work. Their first cost is low but life of light, fast craft is short and engines will be a problem. Consideration should be given to a heavy lift ship as base. The cost might be £350,000 which makes motor boat solution less cheap. Larger motor boats with more endurance and better living conditions might be better. If required could order a prototype for trials, particularly with asdic.¹⁰

Ancestor of the Flower class, the World War I escort Kilbeggan. Some of this class were built by Smith’s Dock.

It is possible that the genesis of the Fairmile A may be discerned in this memorandum.

Options for ‘Cheap’ A/S Vessels¹¹

Conversion of commercial trawler. About 620 tons, 11–12 knots. Coal burning, cylindrical boiler and single reciprocating engine. Endurance about 3,500 at 9 knots. Complement 24. They were moderate asdic platforms but had inadequate subdivision. Conversion would take 4 weeks and cost £35,000.

Admiralty trawler. 510 tons. 11¾–12½ knots. Coal burning, cylindrical boiler and single reciprocating engine. Endurance 3,500 at 9 knots. Complement 24. Good asdic platform with adequate subdivision. They would take 4 months to build and cost £57,000.

Converted whale catcher (Southern Pride). 700 tons, 16 knots. Oil fuel, two boilers, one reciprocating engine. Endurance 4,000 at 12 knots. Complement 30. Moderate asdic platform, subdivision bad. Conversion would take 6 weeks and cost £75,000.

New whale catcher to Admiralty requirements. 900 tons, 16 knots. Oil fuel, two boilers, one reciprocating engine. Endurance 4,000 at 12 knots. Complement 30. Good asdic platform with adequate subdivision. They would take 7 months to build and cost £90,000.

A/S version of Bangor. 500 tons, 17 knots. Oil fuel, two boilers, turbines (alternative diesel). Endurance 4,000 at 10 knots. Complement 50. Good asdic platform and good subdivision. They would take 8 months to build and cost £135,000.

Simplified Guillemot. 580 tons, 20–1 knots. Oil fuel, two boilers, geared turbines. Endurance 3,000 at 11 knots. Complement 63. Good asdic platform and good subdivision. They would take 8 months to build and cost £160,000.

Hunt class. 890 tons, 29 knots. Endurance 3,500 at 20 knots. Oil fuel, two boilers, geared turbines. Complement 144. Good asdic platform and good subdivision. They would take 12 months to build and cost £400,000.

Southern Gem was typical of the big whale catchers built by Smith’s Dock. She is shown in 1942 fitted for A/S duties.

It was noted that all would be ‘seaworthy craft capable of hard work’ but not equal. Similarly they were ‘vessels in which men can live in reasonable conditions’ but again not equal. Protection depended on transverse subdivision and the commercial trawler and whale catcher were ‘very unsatisfactory’ in that regard. Time and cost figures are relative. Note the cost of Guillemot, far larger than usually quoted. Faced with these figures, the Board chose option 4, which developed into the Flower class – and who can blame them?

The problem was still seen, particularly in respect of coastal work on the east coast. The little coastal sloops of the Kingfisher class were capable – and beautiful – but rather shallow for asdic work and far too expensive (Kingfisher cost £160,000) to build in numbers. They displaced 550 tons, coming under a clause of the London Treaty permitting unrestricted building of vessels under 600 tons. Trawlers, particularly those of Admiralty design developed from Basset, were cheap and useful but their speed of twelve and a half knots and short endurance limited their ASW capability, as did their size. Something bigger but still cheap was needed.

It seems that ideas were sought from several builders but details have only survived for the successful candidate from Smith’s Dock. This was a well-known shipbuilder on the Tees specialising in fishing vessels and most notable for its whale catchers. Their managing director, Mr W Reed, pointed out that they had been building A/S vessels since the ‘Zed’ whalers of 1915, followed by the ‘Kil’ class boats, also of World War I. These ‘Kils’ were originally intended to have oil-fired, water-tube boilers for a speed of seventeen to eighteen knots, remarkably similar to the Flowers of World War II. However, oil was scarce, as were skilled personnel, and they completed with coal-fired Scotch-type boilers and a speed of fourteen knots.

Reed’s first proposal in 1938 was based closely on the whale catcher Southern Pride, lengthened by thirty feet. There was a meeting in January 1939 at which Reed seems to have been given some degree of approval for a 700-ton ship costing £90,000. It then grew to 1,390 tons, mainly as a result of a change to coal burning. Fortunately, sanity returned and final approval was for an oil burner of 940 tons (standard). The DNC (Sir Stanley Goodall) was an enthusiastic supporter of the proposal, noting in his diary, ‘I spoke against Guillemot and for whale catcher.’¹² Initially they were known as ‘patrol vessels of whale catcher type’. The origin of the term ‘corvette’ is unclear; it is often said that Churchill chose it and this may well be true, though no evidence has been found to support this. Canadian sources attribute it to Adm Nelles, RCN. Both could be right. Historically it was a very unsuitable name, as a corvette was much bigger than a sloop, but it had a fine ring to it.

The pretty Mallard was a possible alternative to the Flowers but was too expensive and used scarce turbine machinery.

At a meeting on 8 February 1939 Messrs Edwards and Reed of Smith’s Dock drew attention to the performance of the steam trawler Imperialist, which they would guarantee for thirteen knots loaded with 1,050ihp (indicated horsepower) on wet steam.¹³ DNC was not interested, as the extra speed was little more than the Admiralty design for which Smith’s Dock were doing the drawings and the Admiralty design was easier to build. DNC was more interested in the whale catchers Southern Pride and Sondra. The drawbacks to these were poor subdivision and the bar keel. Reed thought he could produce an intermediate design with speed of fifteen to sixteen knots. It was agreed that he should look into the possibility and send an outline drawing and particulars of dimensions, speed and draught, and state time to build, cost and breakdown of equipment between ASI/commercial.

Displacement (tons): 940, 1,170 deep

Dimensions (feet): 190 pp, 205¼ oa x 33 x 13¼ deep

Shp and speed (kts): 2,750ihp = 16 deep

Fuel (tons), endurance (miles) @ (kts): 230, 3,450 @ 12

Complement: 47 (up to 85 later)

Dr Harland has pointed out¹⁴ that the Flowers were far from a copy of the Southern Pride. They had a flat plate keel instead of the whale catchers’ bar keel. Corvettes had bilge keels, inadequate in size at first. They were given a pair of stockless anchors and a windlass. A forecastle was added, forming a seamen’s washplace, heads and stores. The mess deck was below and traditionalists were horrified that seamen and stokers messed together.

The bridge block was sited above the wardroom and two officers’ cabins. On the lower level there was the CO’s cabin and the officers’ bathroom, with a wheelhouse above. On top there was an open bridge with an enclosed compass shelter. In early years there were many individual variations in bridge details but later most were altered to a standard design. The POs’ mess was aft with the galley above, ensuring that food was cold before it reached the forward mess deck.

It was originally thought (1939) that these ships could be used to enforce the blockade on the Northern Patrol. To this end they were given a long-range radio that required two masts some distance apart. Six RN corvettes completed with the two-masted rig but many more retained the foremast ahead of the bridge, where it interfered with the view ahead. All early RCN ships had two masts.

The machinery was little changed from Southern Pride, a four-cylinder, triple-expansion engine driving a single shaft. Such engines were simple to build and within the capability of the engineering departments of most shipyards. Smith’s Dock supplied patterns to other builders and a total of 1,150 units were built for corvettes, frigates and transport ferries (LST 3). Shaft rpm was increased to 185, about the limit for a reciprocating engine lacking forced lubrication. At this speed the engine developed 2,750ihp giving a ship speed of sixteen knots, much faster than any trawler but less than a surfaced U-boat. The machinery was generally reliable, though the maintenance task was heavy. There were early problems with crankshaft alignment, which led to some bearing failures. The majority had two Scotch-type boilers but those from Harland and Wolff had Howden Johnson units, and about twenty later ships (mostly Canadian) had water-tube boilers in closed stokeholds. The particulars of these war-winning engines were: stroke 30in; high-pressure diameter 18½in; medium-pressure 31in; both low-pressure 38½in.

Spirea was an early British Flower. Note the open well deck abaft the forecastle ensuring that the crew got wet on the way to the bridge – and the galley was right aft.

Genista. This picture shows some of the aspects in which the Flowers departed from whaler practice. There was no bar keel, bilge keels were added – and later enlarged – and the bow shape differed. (Goss)

Vibration can be a problem on fast-running reciprocating engines and they were balanced using the Yarrow Tweedy Slick method (based on nineteenth-century torpedo boat work by Schlick). Later, one engine was used for a special vibration trial assisted by Lloyd’s Register. The engine was decoupled from the drive shaft and run up to full speed to identify any critical rpm but no problems were found.

Cowslip. The censor has removed the pendant number but it may be read in the flag hoist (I learnt that trick at school during the war).

On trials, the first ship, Gladiolus, made 16.6 knots with 2,813ihp at 187 rpm on a displacement of 1,118 tons.¹⁵ This equated to 16.5 knots at 2,750ihp and 185 rpm. Fuel consumption was 11.8 tons per 24 hours at 12 knots. Whale catchers need to be able to turn quickly, as do A/S ships using depth charges. The Flowers had the deadwood well cut away and a balanced rudder of 86 square feet. On trials at full speed they had an advance of 2.4 times the ship length and a tactical diameter of 2.1 lengths. Time to turn through 16 and 32 points was 51 and 109½ seconds respectively.

The large number involved, combined with cancellations and transfers, makes certainty over production figures impossible. However, the general picture is clear and is summarised in table 2.3.¹⁶

Flower class corvettes saw extensive service with the RCN.¹⁷ Under the 1939–40 programme, sixty-four ships were ordered in January 1940, the first fourteen to Admiralty specification. Six similar ships were ordered under the 1940–1 programme. Under the same programme, ten Modified Flowers were ordered, with increased sheer and flare, and a long forecastle. The fifteen and twelve ships ordered in 1942–3 and 1943–4 respectively had the same features but more rake to the stem and, most importantly, endurance increased to 7,400 miles at 10 knots.

A total of twenty-two ships were ordered for the French navy. Six were ordered from French yards, of which two were broken up on the stocks after the German invasion. The remaining four were completed as German patrol boats PA1–4. All were sunk and they played no part in the Battle of the Atlantic. Six were ordered from Harland and Wolff and were reordered for the RN as part of the 8 April 1940 order. Of the Smith’s Dock six, one, La Bastiaise, was mined on trials. This was a special tragedy as the engineer manager of Smith’s Dock and several of his staff went down with the ship. Later, in 1941–2, eight British-built Flowers were transferred to the Forces Navales Françaises Libres. Arguments over payment for the French ships continued for a long while.

Clematis. This view shows the open well deck in front of the bridge making passage wet.

Date	Number	Notes
26 July 1939	26	1939–40 Estimates
31 August 1939	30	These two orders went to seventeen small yards not normally warship builders
19 September 1939	20	All Harland & Wolff
21 September 1939	10	Small yards
12 December 1939	10	Small yards
15 December 1939	19	Small yards
8 April 1940	14	Small yards
Rest of 1940	11	Small yards (three had River names originally)
1941	2	One reordered as Castle
1942	7

When the US entered the war in December 1941, they found that the USN was very short of escorts, and ten RN corvettes were transferred as PG62–71. Fifteen more building in Canada were allocated to the USN, though they finally took only eight, numbers between PG86 and PG96, the other seven coming to the RN. In the ex-RN ships a US 4in replaced the British gun, and a 3in the pom-pom. The Canadian-built ships mounted two 3in AA.

Problems, Real and Imaginary

‘Moral is, don’t try and force cheap ships on the Navy, which as Winston says, “always travels first class”.’¹⁸

Trillium shows the appearance of most Flowers by 1942.

Asphodel in 1943 shows the features of most RN early Flowers by that date: the standard bridge with Type 271 radar and Oerlikons.

Mahratta (ex-Charlock), a modified Flower with increased sheer and flare. The funnel is vertical and the conspicuous ventilators of the earlier ships are missing, showing that she has forced-draught boilers.

The entry of the first ships into service from April 1940 brought many complaints. Some were well justified but others arose from the RN’s lack of familiarity with small ships in a seaway. As discussed above, they were intended to have two big masts. The after mast was omitted after six British-built ships, but about fifty Canadian vessels received two masts. The big foremast in front of the bridge was a serious obstruction to the important view ahead. In most, the mast was resited when the bridge was rebuilt in 1941. Gladiolus was the first, completing on 6 April. She was to have a short but exciting life. She sank U-26 on 1 July 1940 and U-65 on 28 April 1941. After several other encounters she was herself sunk, probably by U-553, on 17 October 1941.

The original complement proposed was an unrealistic twenty-nine but by the time they entered the Western Approaches they carried a total of forty-seven. Additional equipment and the need for continuous watchkeeping brought further increases, and by January 1941 Asphodel had a complement of sixty-seven men. The first fifty ships had no thermal insulation on the ship’s side and this, combined with poor ventilation, led to heavy condensation, which together with leaky seams contributed to a high incidence of tuberculosis. Seven POs shared a toilet and three washbasins. The ratings’ quarters measured about twenty-two feet by thirty feet for fifty-five men, with a single urinal, two toilets and four wash bowls. Sleeping quarters were in two compartments of twenty feet by fourteen feet on the lower deck, making escape difficult (see escape, chapter 8).

Trentonian, a Canadian modified Flower.

Arbutus, a modified Flower; note the raised Hedgehog.

There was a single, small galley at the after end of the superstructure and all food had to be carried across the open well deck, frequently swept by heavy seas, to the forecastle. There was a single, small refrigerator. Fresh food would be gone in about five days, after which only tinned supplies were available. Seasickness reduced the need for eating.

A considerable improvement was made by extending the forecastle to amidships in 1941–2. This helped to keep the deck dry and made extra space available for accommodation. There was a slight offset in that the extra weight forward reduced the freeboard.

It is a truism that small ships are more lively in a seaway than are their bigger sisters; indeed more recent work has shown that seasickness is almost a function of the length of the vessel. The original Flowers rolled badly, partly due to inadequate bilge keels (see table 2.4). Rolling particularly affects manual tasks such as reloading depth charge throwers and other weapons. A frequent comment was that a Flower would roll on wet grass. Bigger bilge keels helped to some extent but these small ships when encountering big seas were forced to roll in the period at which they encountered the waves.

Bridge layout was, and still is, a matter of considerable debate amongst seamen, with varied and often conflicting views. The one thing that RN officers were agreed on was that the bridge should be open and, with 30–50 per cent of first sightings made visually up to the end of 1942, they were surely right (though the USN preferred enclosed bridges). It was generally believed at the time that plenty of cold, fresh air helped to keep lookouts and the officer of the watch awake. Post-war research shows that cold is likely to reduce alertness.

The original bridge design incorporated an enclosed wheelhouse, which also contained the asdic display. Changes were soon made, at first on an ad hoc basis but the introduction of radar meant that another office was needed, close to the aerial. It was also decided to fit machine guns in the bridge wings, later changed to single Oerlikons, which had to be extended to give the guns a decent arc of fire and to keep their moving, ferrous mass away from the magnetic compass. This led to a standardised bridge design fitted in almost all ships at the same time as the forecastle was extended. The package update, comprising bridge and forecastle extensions, Hedgehog and asdic 144, took about fifteen weeks.

At the front of the bridge there was an overhanging box containing the asdic and a small chart room. The compass was raised behind. The bridge was the full width of the ship, with a single Oerlikon either side (later replaced by a 6pdr in some ships). At the back was the radar office for the Type 271, with its lantern above. A few ships had this on the centreline but in most it was offset to port to improve the view aft. The wheelhouse was found to obstruct the view for visual signalling, so important before TBS, and the wheel was moved below. A few ships had a similar bridge adapted for a short forecastle.

The 4in gun was mounted on the forecastle on a low platform. Even this allowed insufficient depression for close range shots over the bow and the gun was raised further in the Modified Flowers, which involved raising the bridge. Later in the war, two more single Oerlikons were added each side in the waist. A few ships had the pom-pom replaced by a power-operated twin Oerlikon.

Some corvettes had radar 286M (fixed) or 286P (rotating). These were useful for station-keeping but not very effective against submarines. Early Canadian ships had SW-1 and 2, of generally similar performance. At the end of March 1941 the first 271 set underwent trials in Orchis. This was a giant leap forward, particularly when fitted with PPI.

Initially, Flowers had twenty-five depth charges with two rails and two throwers. This was soon increased to forty charges, and later, two more throwers were added. The original asdic was type 123 (trawler dome, ie non-retractable) In 1942–3 a split Hedgehog was fitted behind the 4in gun, with twelve spigots either side of the ship.

In later ships, the hull form was modified, with increased flare and sheer. The 4in gun was raised to fire at depression over the raised bow, which, in turn, meant raising the bridge by one deck. Most, though not all, of these Modified Flower class had a short mainmast. They had water-tube boilers with forced draught through a closed stokehold. This eliminated the ‘trademark’ four cowl ventilators of the earlier ships. The funnel was vertical with a cinder screen (on most) to make it more difficult to determine the course. There were twenty-two UK built ships and about thirty-seven in Canada. There were some intermediate ships with some of these features only. The Flower class ships could be built on the Great Lakes.

Burnet, a modified Flower, unusual in having a tripod mast.

Modifications to RCN ships followed the same pattern as the RN’s, but distance from the supplier usually meant that modifications came much later in the Canadian ships.

The A–I Class Destroyers – Excellence to Obsolescence

At the end of the first world war the V&W classes of destroyers were the best destroyers in the world, at least for a North Sea war: seaworthy, well armed and with adequate endurance. There were a considerable number of these ships available – more than enough for the peacetime fleet – so there was no urgency in the need for new designs.²⁰

By November 1923, agreement had been reached on outline requirements for a new destroyer; it was very much a ‘W Mk II’. DNC’s destroyer section under C J W Hopkins prepared three sketch designs around these requirements and these were shown to five leading destroyer builders, who were invited to use their expertise in developing the best future destroyer. Designs by Thornycroft and Yarrow were preferred and orders were placed with them for Ambuscade (Yarrow) and Amazon (Thornycroft). As usual, Yarrow had gone for light weight, whilst Thornycroft’s design was bigger and more powerful.

In the late summer of 1926, before the two prototypes completed, requirements were drawn up for the A class of eight destroyers and a rather bigger leader. They were designed by DNC’s destroyer section, though some features of the Thornycroft design were incorporated. These destroyers were generally well liked, though there was a widespread view in the Royal Navy that they were too big, too complicated and too expensive, while at the same time much extra equipment was proposed. Many similar ships followed, with a total of seventy-nine ships for the RN and RCN up to the I class. In addition, at the outbreak of war there were eight similar ships building in the UK for other countries (six H class for Brazil and two I class for Turkey), which were taken over.

Last of the line: HMCS Sackville, preserved at Halifax, Nova Scotia.

	A class	I class
Displacement (tons):	1,330, 1,738 deep	1,450, 1,877 deep (typical)
Dimensions (feet):	323 x 32¼ x 10¼	323 x 33 x 12½
Shp and speed (kts):	34,000 = 35.25	34,000 = 36
Fuel (tons), endurance (miles) @ (kts):	390, 4,650 @ 15	450, 5,300 @ 15
Complement:	138	145
Cost:	~ £250,000	~ £250,000

All these vessels were planned as fleet destroyers, with torpedo attack on the enemy battle fleet as their primary task, together with frustrating attacks by enemy destroyers on their own fleet. Initially, the intention was that alternate flotillas would have asdic, and the others would have the two-speed destroyer sweep (TSDS) instead. The As were intended to have asdic but there was no production set available and they completed without, although they did receive their asdic sets before the war. The Bs were fitted on build, and from the D class onwards asdic was fitted to all during build. It was said that fitting asdic added £2,500 to the quarter of a million pounds that the vessels cost.

Even in their designed role, these ships had a number of weaknesses, most of which were recognised at the time but could not be cured within the budget or technology available. There was a demand for greater endurance, met in part by more oil, but this made the ship bigger and more expensive. The machinery of the new ships was more efficient than that of the wartime destroyers, mainly as a result of the use of superheaters, but also because of better materials. A further improvement was attempted in the Acheron with the use of steam at a much higher temperature and pressure (700ºF and 500lb/in² in Acheron, 300lb/in² in the rest; cf the contemporary USS Mahan, 700º, 400lb/in²). This plant in Acheron was successful in improving efficiency, with a fuel consumption of 0.608lb/shp/hr compared with a figure of 0.81 for her sisters. However, Acheron’s machinery gave a lot of trouble, which was never cured. The main problem lay in vibration of the blades in the impulse stage of her Parsons turbines. US sources say that the Parsons design team of this era were very old fashioned and had neglected recent work on blade flutter.²¹ By the late 1930s, USN machinery was some 20–30 per cent better in fuel economy, lighter and more compact, and required much less maintenance.

Brilliant in 1943 with a typical destroyer fit for that date. She has lost Y gun but has Hedgehog, extra depth charges, radar, HF/DF and Oerlikons.

There were a few experimental boilers but no significant improvement was shown. All these ships had three boilers, so that they could keep up with the fleet using two only while one was being cleaned or otherwise maintained. By the outbreak of war, RN boilers needed cleaning every 750 running hours (cf USN, 2,000 hours), which fitted in well with the peacetime leave routine. In the earlier ships, the three boilers were in two rooms: two boilers in the forward space, one aft. This meant that it was easy to immobilise the ship with a hit anywhere in the machinery.²² There was a slight improvement from the E class onwards, with three boiler rooms. A unit system (ie with two engine rooms) would have made the ship ten feet longer and cost an extra £15,000.

Another well-known problem was that the 4.7in guns were low angle only and had no capability against closing aircraft. Many attempts were made to give at least one gun a 60º elevation but no success was achieved. The problem was that high angles of elevation needed high trunnions, but this would make the gun very difficult to load at low angles. It may be concluded that a hand-loaded, dual-purpose gun was not feasible and that power operation was needed. Since the RAF advised that dive-bombing was not a serious threat and that high-level bombers could not hit a manoeuvring destroyer, the lack of an AA gun was not seen as too serious. The low-angle guns could protect the battle line in barrage fire, though the RN AA control system was poor.

The ships were constructed of D-quality steel, riveted, with transverse frames. Stresses were reasonable for this style of construction. The single-riveted seams, particularly on the forecastle deck, were very prone to leaking. The weight saving achieved by the use of D-quality was not great; this was studied in the design of the Hunt class, where a saving of thirteen tons was estimated. D-quality was difficult to weld and other navies that welded similar steels had many problems. However, it is probable that a bigger weight saving would have been obtained by adopting a welded, mild-steel structure. Some welding was introduced with the H class.

	Hogging (tons/in2)	Sagging (tons/in2)
Keel	6.2	5.7
Deck	8.3	5.5

Intact stability was satisfactory, with a GM of 2.4 feet (light), 2.25 feet (deep) in the I class. They were expected to have a slight positive GM with the engine room and adjacent boiler room flooded. This is far below modern standards but was adequate, since only six capsized after a single torpedo hit; twenty-two broke in half. None were lost to the stress of weather alone. For some reason unknown, Duncan had two boilers in the room next to the engine room, making her liable to a great flooded length if hit in this region.

By 1941, weight growth was a serious problem, with 100–150 tons additional weight for radar, HF/DF, Oerlikons, splinter protection and depth charges – the heavy Mk VII was a particular problem. Too many coats of paint were another load, particularly on the leaders, which always looked immaculate. Some three to four inches of metacentric height had been lost out of an initial thirty inches in the light condition. Most had a fuel restriction imposed, requiring them to take in water ballast if fuel fell below forty tons – a further restriction on endurance. This restriction was generally ignored to save the effort needed to clean the tanks before refuelling.

The initial main armament was eight 21in torpedo tubes in quadruple mounts (ten in the I class). In the real war, these were often known as the ‘main ornament’. There were four 4.7in QF guns (five in most leaders) and two single pompoms. They had two depth charge throwers and four chutes with, initially, eight charges (increased to fifteen by the outbreak of war). They all had asdic 128 by 1939.

During the war A–I class ships were gradually replaced on fleet duties by later vessels and relegated to ‘secondary’ duties such as ASW. The A and B classes were already available for such work at the start of the war. Initial changes were few; the after set of torpedo tubes was replaced by a 3in AA gun, which, without control, was useless. The number of depth charges was increased with, in most cases, the loss of Y gun. Oerlikons were added as they became available, with six in most survivors by 1944.

Some ships were given a more specific ASW fit, with A gun replaced by Hedgehog with asdic 144 plus Q (see appendix IV). The number of depth charges steadily increased, with 35 in early war years rising to between 60 and 135 in some ships with four throwers. Ambuscade was a test bed for Parsnip²³ and later for Squid. Escapade had a serious explosion in her Hedgehog and was given Squid during repairs.

As escorts, their virtue was their speed and they were often used to pursue a contact, accounting for a high proportion of ‘kills’ in the early years. On the other hand, their endurance was insufficient to escort a convoy across the Atlantic. Refuelling at sea was only a partial answer, as the techniques and equipment available were time-consuming and could be used only in good weather. Some consideration was given in September 1943 to a long-range escort conversion. This would have been difficult, as almost all had two boilers in the forward room, which would have meant an extra oil-tight bulkhead for the new tank.

It is interesting that the conversion package would have included strengthening by doublers on the upper deck and thicker garboard strakes – incidentally reducing the ballast needed! The overall strength of these ships appears to have been satisfactory but wave impacts and slamming led to leaky single-riveted seams, which were unpleasant in mess decks and potentially disastrous in feed-water tanks. The break of forecastle amidships formed a structural discontinuity, which led many to break in half after damage.

Their low freeboard to the upper deck made its passage hazardous in bad weather and reloading depth charges was particularly dangerous. Their turning circle was on the wide side for depth charge attacks but overall they were effective ASW vessels of their day, participating in thirty-nine sinkings.

Hurricane as leader of B1 Escort Group with a similar ASW fit to Brilliant.